African Easterly Jet:Structure and Maintenance

Man-Li C. Wu*

Global Modeling and Assimilation OfficeNASA Goddard Space Flight Center

Greenbelt, Maryland 20771

Oreste Reale tLaboratory for Atmospheres

NASA Goddard Space Flight CenterGreenbelt, Maryland 20771

Siegfried D. SchubertMax J. Suarez

Randy D. KosterPhilip J. Pegion t

Global Modeling and Assimilation OfficeNASA Goddard Space Flight Center

Greenbelt, Maryland 20771

May 20, 2009

*Email Man-Li.C.Wu@nasa.govtAdditional Affiliation: University of Maryland, Baltimore County, Baltimore, Maryland# Current Affiliation: National Ocean and Atmosphere Administration, Camps Springs, Mary-

land

1

Abstract

This article investigates the African Easterly Jet (AEJ), its structure and the forc-

ings contributing to its maintenance, critically revisiting previous work which at-

tributed the maintenance of the jet to soil moisture gradients over tropical Africa.

A state-of-the-art global model in a high-end computer framework is used

to produce a 3-member 73-year ensemble run forced by observed SST to repre-

sent the Control run. The AEJ as produced by the Control is compared with the

representation of the AEJ in the European Center for Medium Range Forecast

Reanalyses (ERA-40) and other observational data sets and found very realistic.

Five Experiments are then performed, each represented by sets of 3-member

22 year long (1980-2001) ensemble runs. The goal of the Experiments is to inves-

tigate the role of meridional soil moisture gradients, different land surface proper-

ties and orography. Unlike previous studies, which have suppressed soil moisture

gradients within a highly idealized framework (i.e., the so-called bucket model),

terrestrial evaporation control is here achieved with a highly sophisticated land-

surface treatment and with an extensively tested and complex methodology. The

results show that the AEJ is suppressed by a combination of absence of meridional

evaporation gradients over Africa and constant vegetation, even if the individual

forcings taken separately do not lead to the AEJ disappearance, but only its modifi-

cation. Moreover, the suppression of orography also leads to a different circulation

in which there is no AEJ. This work suggests that it is not just soil moisture gradi-

ents, but a unique combination of geographical features present only in northern

tropical Africa, which causes and maintains the jet.

1. Introduction

The African Easterly Jet (AEJ) is one of the most complex and intriguing dynami-

cal features in tropical meteorology. It raises a number of questions, ranging from

its formation mechanism to its maintenance, and from the causes of its intensity

fluctuations to its role in generating weather systems. The AEJ is a crucial element

in global-, synoptic- and meso-scale dynamics, and its representation in models is

important for climate modeling, seasonal predictions, and weather forecasting.

This work uses a combination of state-of-the-art global reanalyses and model

experiments, to shed light on some aspects of the AEJ. A general discussion of

the problem and a literature review are presented in the next Section. Section

3 provides a description of the data and the NASA Seasonal-Interannual Predic-

tion Project (NSIPP) Atmospheric General Circulation Model (AGCM), version

1 (NSIPP-1). Section 4 describes a set of five Experiments and a Control run that

are carried out to investigate the role of soil moisture, vegetation and orography

in maintaining and controlling the AEJ. In Section 5 the structure of the AEJ is

analyzed as it appears in the reanalyses and in the Control run. This is followed

by a discussion of all the forcings introduced in the Experiments and the resulting

impact on the AEJ. Section 6 discusses the dynamical implications of the Experi-

ments and Section 7 states the conclusions of this work.

2. The AEJ: previous research

Besides the Inter Tropical Convergence Zone (ITCZ) the three planetary features

that most prominently control the northern African summer and tropical Atlantic

weather systems are the Saharan thermal low, the Tropical Easterly Jet (TEJ) and

the AEJ (e.g. Asnani 2005). The Saharan thermal low can be considered as an

extension of the planetary-scale thermal trough associated with the Asian mon-

soon. It affects the troposphere from the surface up to about 600 hPa above which

is replaced by relatively high geopotential values and divergence. It is the main

feature controlling the so-called Harmattan flow, a low-level predominantly north-

easterly flow sweeping over northwestern Sahara during summer and contributing

to Saharan dust outbreaks into the Atlantic ocean. The TEJ, located at about 100-

200 hPa, 5° to 15°N, and stretching from Indo-China to the tropical Atlantic, has

been shown to be linked to anomalous SSTs on a planetary scale (Chen and Loon

1987). It plays a paramount role in controlling the easterly vertical shear over

developing weather systems off the Atlantic coast of Africa.

It appears, however, that of the above phenomena, the AEJ exerts the strongest

influence on the occurrence, intensity, and frequency characteristics of the African

weather systems within the west African Monsoon region (e.g. Cook 1999). That

influence appears to extend well beyond the monsoon region to impact a substan-

tial fraction of tropical Atlantic weather systems (e.g. Asnani 2005).

Hamilton and Archibald (1945) should be remembered among the first authors

to present evidence of the AEJ. Based on a surface wind-based definition of the

2

ITCZ (i.e. wind-veering line) they described Sahelian Africa to be characterized

by four fundamental zones. Regardless of the fact that modern definitions of the

ITCZ are based on the line where maximum vertical velocities occur rather than

as the surface wind-veering line, the distinction of those four environments (A:

north of the wind-changing line, characterized by dry Saharan air, B: between the

AEJ and the wind-changing line, C: south of the AEJ, and D: far away to the

south of the AEJ) and their different weather patterns, is still of enormous value

and emphasizes the importance of the AEJ position. Moreover, these authors

correctly ascertained the important role of the AEJ as a triggering mechanism of

weather phenomena and the confinement of the weather systems to a very narrow

band to the south of the AEJ.

In particular, the zone to the immediate south of the AEJ core, being subject

to cyclonic shear and strong convective activity, has been recognized for decades

as the area in which African Easterly Waves (AEWs) form, where several weather

disturbances develop in form of squalls, and where the maximum precipitation is

recorded. In particular, some authors have separated squall- by nonsquall- AEWs

(e.g. Dhonneur 1971) and other authors have provided a much deeper understand-

ing of squalls (e.g. Houze 1977; Zipser and Gautier 1978). To this day, one of the

most comprehensive studies on AEWs is perhaps the 3-part work by Kiladis et al.

(2006); Hall et al. (2006); and Thorncroft et al. (2008).

However, in this work AEWs are not treated and the focus is only on the AEJ

structure and maintenance. Despite the importance of the AEJ as a prominent fea-

ture of African meteororology, the observational network over northern Africa has

3

been deemed not adequate to completely capture its fine structure. Parker et al.

(2008) state that “in recent years the meteorological observing system of Africa

has been in a worse state than that of any other populated continent.” In addition

to the poor observing network, the diagnosis of the AEJ through global, relatively

low-resolution datasets, has shown some degree of uncertainty. Differences in

height, shape and intensity of the jet were noted by Cook (1999), while compar-

ing the gridded climatologies of the National Centers for Environmental Predic-

tions (NCEP) with the European Centre for Medium-Range Weather Forecasts

(ECMWF). With respect to the AEJ, Cook (1999) notes that the AEJ maximum

easterly velocities are 1 1m s-1 at 500 hPa in the ECMWF data, and 12 . 5m s-1 at

600 hPa in the NCEP reanalyses (Kistler et al. 2001), contrasting with a 8-year

observational data set (Burpee 1972) that shows the AEJ max at about 650 hPa.

Moreover, as observed by Cook in the same work, the shape of the AEJ differs too,

being more displaced to the east in the NCEP data. These important differences

were attributed by the author to sampling problems and relatively low resolution.

The relative scarcity of observations is also stressed in a more recent observa-

tional experiment, called Jet 2000 (Thorncroft et al. 2003), in which the jet height

was found at about 675 hPa, and more to the south than expected, but possibly

in agreement with the behavior of the AEJ in dry years. The Data Assimilation

and forecasting Systems from ECMWF and the United Kingdom Met Office were

initialized with some of the campaign data and provided reasonable representa-

tions in the AEJ although some differences between the two systems were noted

(Thorncroft et al. 2003).

4

Increasing vertical and horizontal resolution in more recent data sets have not

yet totally dissipated some ambiguity in the AEJ definition, as it will be shown

later. In addition to insufficient observational coverage and models’ resolution,

several important aspects may be at this point still inadequately represented. For

example, Tompkins et al. (2005) show that the ECMWF 5-day forecast of the

AEJ, which had a bias with respect to observations (being too far north, too weak

and at a greater height than observed) was substantially improved by including a

realistic, annually-varying, aerosol climatology. A different treatment of aerosols

may be then at least partly responsible for the discrepancy.

Given that uncertainties in defining the strength, horizontal extent and height

of the observed AEJ may be reduced in the future by a denser observational net-

work, as advocated by the proponents of the African Monsoon Multidisciplinary

Analysis (AMMA, described, among others, by Redelsperger et al. 2006), im-

proved data assimilation capabilities and higher resolution, this will not directly

address the more fundamental problems related to the AEJ, involving its physical

causes, its maintenance and its ability to produce and confine the tracks of weather

systems.

A relevant study that assesses the possible forcings on the maintenance of

the AEJ is by Thorncroft and Blackburn (1999). In this work the authors com-

pare the importance of the surface fluxes, low-level heating and dry convection

associated with the Saharan thermal low, against the role of deep moist convec-

tion taking place in the ITCZ. Their conclusion is that it is the low-level heating,

which extends vertically up to about 700 hPa, to play the most prominent role in

5

maintaining the AEJ.

Accepting that the mean position of the AEJ may be controlled predominantly

by low-level thermal forcing, the problem of interannual variability connected

with intensity fluctuations or changes in position has been the subject of investi-

gation for several decades. Amidst the considerable amount of literature available,

perhaps two different but not necessarily antithetical views can be extracted: one

that emphasizes the importance of global or Atlantic or Indian Ocean SSTs, and

one that emphasizes the role of land-surface properties. In the former category,

one can mention among many other studies the work by Foland et al. (1986);

Semazzi et al. (1996); Nicholson (1997); Nicholson and Kim (1997), and Vizy

and Cook (2001). In the latter category, studies stressing the importance of land

surface properties start with the seminal works of Charney on the albedo impact

in semi-arid regions (e.g. Charney et al. 1977) and more recently include the

increasingly sophisticated description of land-surface characteristics, including

vegetation properties, as potential modulators of north-African rainfall on a vari-

ety of time-scales (e.g., Xue and Shukla 1993; Zeng and Neelin 2000; Clark et al.

2004; Douville et al. 2005; Hagos and Cook 2005).

All these different perspectives are perceived as fundamentally concurring

mechanisms in the study on Sahelian droughts by Giannini et al. (2003), where ev-

idence is presented that Sahelian rainfall variability may be controlled by oceanic

forcings and amplified by land-atmosphere interaction.

Of particular relevance for the present work is the study by Cook (1999),

which showed with great clarity the importance of the AEJ in controlling west

6

African precipitation, and quantified the connection between the Saharan thermal

low and the steep soil moisture gradients observed from the wet areas close to the

Guinea coast to the dry environment of the Saharan desert, as prominent mecha-

nism controlling the AEJ. Hsieh and Cook (2005) expanded on these findings by

showing the interdependence of the ITCZ and the AEJ in the generation of AEWs,

Cook and Vizy (2006) analyze and review the capability of coupled GCMs to re-

produce the overall west African Monsoon dynamics, and Hsieh and Cook (2007)

investigate the energetics of AEWs within a regional framework. All these latter

study, and many others, make assumptions based upon Cook (1999) and do not

discuss the maintenance of the AEJ.

In the work presented here, which is focused on the AEJ and not on the AEWs,

we follow Cook (1999) by looking in more detail, and with a less idealized model,

at the impacts of surface conditions on the AEJ. Cook’s study, for example, used

zonally prescribed clouds, perpetual July conditions, a bucket model of the land

surface, no topography and a uniform albedo of 0.1. The conclusion that under

these conditions the AEJ is produced by prescribing soil moisture alone does not

address the importance of other factors or, indeed, the details of the soil moisture

distribution or its relative impact in a more realistic setting. We thus decided to

revisit the problem within a more realistic modeling framework and with a more

rigorous set of experiments.

In this study, we start from a perspective opposite to Cook (1999): after exam-

ining the structure of the AEJ as represented in ERA-40, and in other reanalysis,

we carry out a number of sensitivity experiments with a state-of-the-art global

7

AGCM, with realistic topography and sophisticated land surface, that is capable

of producing a realistic AEJ. In particular, a set of five experiments is performed:

in four of them one forcing at a time is removed and in one two forcings are

removed. By doing so, we separate the roles of soil moisture, orography and veg-

etation. The well-tested strategy developed by Koster and Suarez (1995, hereafter

KS95), and Koster et al. (2000, hereafter KA00) is used to control land evapora-

tion in AGCMs (see also Reale and Dirmeyer 2002; and Reale et al. 2002).

3. Data and Model description

The data used for this study are: a) the ERA-40 reanalyses (Uppala et al. 2005), a

6-hourly global data set covering more than 40 years, produced by the ECMWF,

describing the state of the atmosphere from September 1957 to August 2002; b)

the NCEP Department of Energy (NCEP-DOE) Reanalyses 2 (NCEP R2, Kistler

et al. 2001; Kanamitsu et al. 2002) which cover about sixty years; and c) the

Japanese 25-year Reanalyses (JRA-25, Onogi et al., 2007). In this work, only the

subperiod between 1980 and 2001 is being used.

The model is the Global Modeling and Assimilation Office (GMAO) NSIPP-

1 AGCM, run at a horizontal resolution of 20 x 2.50 and with 34 vertical levels

spaced to provide higher vertical resolution in the lower 2 kms of the atmosphere.

It is the same model that has been previously used by Wu et al. (2006) to investi-

gate the seasonality and meridional propagation of the Madden Julian Oscillation

(MJO). NSIPP-1 is a grid point model using a 4th-order dynamical core (Suarez

and Takacs, 1995) with explicit leapfrog time differencing. The vertical coordi-

8

nate is a standard sigma-coordinate and the vertical differencing follows Arakawa

and Suarez (1983). The boundary layer scheme is a K-scheme calculating tur-

bulent diffusivities for heat and momentum based on Monin-Obukhov similarity

theory (Louis et al. 1982). Turbulent diffusivities are determined as function of

roughness length, von Karman constant and a bulk boundary-layer Richardson

number.

The relaxed Arakawa-Schubert (RAS) scheme is adopted to parameterize con-

vection (Moorthi and Suarez 1992). RAS uses a sequence of simple linearly-

entraining plumes (clouds types) that originate at the lowest sigma level and de-

train at specific model levels. The initial clouds-base mass flux for each cloud

type is determined from a convective available potential energy closure. The RAS

scheme effectively acts as a parameterization of both deep and shallow convec-

tion in the NSIPP-1 model. Bacmeister and Suarez (2002) show that the model

produces a reasonable simulation of the thermodynamic structure of the lower tro-

posphere. The adopted parameterization of solar and infrared radiative heating is

described in Chou and Suarez (1994, 1999). The solar heating includes absorption

of O3, CO2, water vapor, O2, clouds, and gaseous and aerosol scattering.

The land surface model (LSM) is the Mosaic scheme of Koster and Suarez

(1992, 1996), a soil-vegetation-atmosphere transfer LSM. The Mosaic scheme

subdivides each AGCM grid square into subregions, called ‘tiles’, in which the

vegetation type is relatively homogeneous, and then calculates separate one-dimensional

energy and water balances over each tile, with strong stomatal control over tran-

spiration rates. The LSM performance against observations has been discussed in

9

Chen et al. (1997) and Wood et al. (1998).

The climate of the NSIPP-1 model and a more detailed model formulation are

described in Bacmeister et al. (2000). Aspects of the model’s climate variability

are assessed by, among others, Pegion et al. (2000) and Schubert et al. (2002).

This is the same model used in the investigation of Sahelian rainfall variability by

Giannini et al. (2003) Of particular importance for this work is the model’s ability

to produce realistic AEJ and the good performance of the LSM documented in the

studies above (see also Figs. 1 and 5 in this study), which makes it suitable for

investigating the role of soil moisture in maintaining the AEJ.

4. Experiments

The Control runs consist of a three-member 73-year ensemble of simulations car-

ried out at a horizontal resolution of 2° x 2.50 and forced with observed monthly

varying SSTs from 1930 to 2002. The atmospheric initial conditions for the three

members of the ensemble were arbitrarily chosen from previously completed sim-

ulations. Unlike earlier work (e.g. Thorncroft and Blackburn, 1999; Cook, 1999),

which used highly idealized models, it should be stressed that this Control run is

a relatively realistic climate simulation.

The purpose of the five Experiment runs is to examine the impacts of modi-

fied surface conditions in Equatorial and Saharan Africa obtained by 1) suppress-

ing directly the meridional gradients of evaporation, 2) imposing meridionally

uniform land surface properties (with two different settings), 3) simultaneously

suppressing evaporation gradients and prescribing land surface properties, and 4)

10

suppressing the orography’s role. Each experiment consists of a three-member

27-year ensemble of simulations (performed at the same resolution as that of the

Control, with the same atmospheric initialization and the same observed monthly

varying SSTs from 1978 to 2004), in which one of the forcings is disabled. The

overall period (1978-2004) was chosen due to the much better quality of reanal-

yses available during the satellite era. The first two years of the each simulation

are discarded for spin-up, and also the last three years are not included, in order to

overlap with the ERA-40. All the statistics and subsequent analysis of this work

are therefore based on the 22-year period 1980-2001.

One important aspect of our experiment settings is that the Control may be

able to represent possible trends or fluctuations connected with changes in land

surface properties. These have been documented over the region, and the actual

Sahara-Sahel border has been shown to be subject to fluctuations (e.g. Tucker et

al. 1991; Giannini et al. 2003).

The land interaction is disabled following the same conceptual design adopted

by KS95 and KA00. The main goal is to disable the impact of evaporation vari-

ability, which is more difficult than disabling SSTs due to the intrinsic differ-

ence between SST’s and soil moisture’s control of evaporation. Following KS95

and KA00, the concept of “evaporability” or “evaporation efficiency” is therefore

adopted. The evaporability 0 is defined as the ratio between actual evaporation

and potential evaporation. The former is given by

E = pE es — ea

(1)ps ?-aer + ?-surf

11

where ρ is the air density, ǫ is constant, ps is the surface pressure, es is the saturated

vapor pressure at the surface temperature, ea is the vapor pressure in the free air

(namely, the lowest σ level), raer is the aerodynamic resistance and rsur f is the

surface resistance, computed as a parallel resistance of all the resistances that

contribute to it. Potential evaporation Ep is calculated from actual evaporation

when the surface resistance rsur f is set to zero.

E = ρǫ es — ea (2)pps raer

For each month and each year of the control run we compute the following ratio

_ P E(i)βmy — E Ep (i)

where the sums are over all the timesteps i of that particular month and year. This

allows the definition of a time series of 22 x 12 monthly βs from the control run

for each gridbox.

The climatological seasonal cycle of evaporability is obtained by averaging β

for each month throughout the 22-year period. Prescribing the evaporation effi-

ciency in a simulation allows us to control continental evaporation in the simula-

tions without directly prescribing the evaporation and yet disabling land-atmosphere

feedbacks due to changes in surface water. As argued by KS95, this is a more gen-

eral and direct way of accomplishing this than attempting to prescribe the water in

all surface reservoirs. Once β is prescribed, the actual evaporation for each time

step is computed by multiplying Ep (i) by the corresponding βm .

In summary, the following experiments were performed:

(3)

12

1. Experiment 1: the climatological set of βms is averaged over the Sahel re-

gion (hereby defined as 101 to 201N and 201W to 351E) and imposed on the

rest of continental Africa. By doing so, evaporation interannual variability

is suppressed and Sahelian conditions are prescribed.

2. Experiment 2: conditions of βm computed from Saharan values are imposed

on the rest of Africa. This experiment acts in the opposite direction of Ex-

periment 1. In fact, Experiment 1 imposes some evaporation over desert,

thus reducing the meridional gradients of evaporation by increasing evapo-

ration to the north of the gradient maximum, whereas Experiment 2 reduces

evaporation over areas which are not desert, thus reducing the gradients by

constraining evaporation to the south of the gradient maximum.

3. Experiment 3: the vegetation is imposed to be uniform, namely the pre-

dominant Sahelian land type (i.e., grassland) is imposed over the entire

African continent, and the subgrid vegetation variability within each tile is

suppressed. The primary effect is believed to be the suppression of albedo

gradients and their effects on temperatures, but it also results in changes

in surface roughness and hydrological properties associated with vegetated

surfaces.

4. Experiment 4: the forcings of Experiment 1 and 3 are combined: vegeta-

tion and the evaporation efficiency are simultaneously prescribed to Sahel

values.

5. Experiment 5: orography is imposed to be flat, eliminating any possible

13

contributions induced by the southward sloping terrain over the Sahel, and

by the Ethiopian highlands.

The Experiments’ settings are summarized in Table 1.

5. Analyses and experiment results

Daily means were extracted from the five Experiments and the Control runs from

the 1980-2001 period. In Fig 1, the zonal wind climatology for July, based on

the ERA-40 data, the NCEP R2, and the JRA-25 are compared with the NSIPP-1

Control Run.

The fundamental properties of the AEJ and of the general circulation over

northern Africa can be clearly distinguished in the 3 sets of reanalyses:

• a maximum of about 11 ms- ' at 600hPa approximately 15°N over western

Africa

• a low-level westerly flow confined below 800hPa and affecting the latitude

band between the Equator and 20°

• a low-level easterly flow (Harmattan) confined between 20° and 30°N

• an upper level easterly flow (TEJ) at latitudes south of 10°N

• a shift in the AEJ position between May and July from about 9°N to 15°N,

followed by a 20% AEJ speed decline in October and southward symmetric

displacement (not shown)

14

Previous work by Cook (1999, Fig. 1) showed a remarkable difference in the AEJ

maximum speed, of about 20%, between the older European Center and NCEP-

NCAR reanalyses, and suggested resolution as a possible cause for the discrep-

ancy. In the more recent datasets, ERA-40, NCEP R2 and JRA-25, wich are

presented in this paper, the AEJ Maximum is at about the same value of approx-

imately 11 ms-i . Also the location of the maximum, which Cook (1999) noted

to differ substantially, tends to be in all analyses concentrated at about 10° — 15°

longitude W.

However, the structure of the AEJ as represented in the two reanalyses still

differs especially on its eastern side, with the ERA-40 representation of the AEJ

consisting of a secondary speed maximum over the Aethopian highlands. The

JRA-25, which is a more recent product, agrees more closely to the NCEP R2

as for the location of the jet maximum, but more to the ERA-40 as for the shape

of the AEJ over central-eastern Africa. Quantitatively, speed differences of about

20% exist between JRA-25 and NCEP R2 at longitudes between 20° and 40°E .

The cross-section at 0° longitude, extracted from the ERA-40, shows a well-

defined AEJ core at about 600 hPa, well separated from the TEJ above 300 hPa and

from the low-level Harmattan flow centered between 27° and 30°N. Comparing

this Figure with Fig. 2 in Cook (1999) and with NCEP R2 there is a good agree-

ment on the low-level monsoonal flow between approximately 2°N and 20°N and

the representation of the AEJ core at about 600 hPa. However, there is substantial

difference in the representation of the Harmattan flow at about 27°N: ERA-40

shows a much deeper Harmattan flow (up to 700 hPa) with a maximum located

15

about 21 in latitude further to the north with respect to the NCEP R2. Considering

the importance of the Harmattan flow in advecting Saharan Air, and the impli-

cations of that for suppressing or limiting convection, the difference between the

analysis seems significant. Another important difference is the intensity of the

TEJ, which affects easterly vertical shear on the equatorward flank of the AEJ.

The TEJ is almost 20% weaker in the ERA-40 data than in the NCEP R2. The

JRA-25 is overall closer to the NCEP R2, except that the AEJ is definitely weaker

on its eastern side in the JRA-25: at about 25°E the difference is of the order

of 40%. Moreover, in the JRA-25 a separate speed maximum is evident on the

Ethiopian highlands. This feature is also present in the ERA-40, but not in the

NCEP R2. At this time, because of the relatively scarce observational network,

these differences can only be acknowledged and it is not possible to make a final

statement on which data set provides a more realistic representation. It should be

emphasized however that even wind speed differences of small magnitude, rela-

tive to a feature as important as the AEJ, may be relevant considering that NCEP

R2, ERA-40 and JRA-25 are state-of-the-art reanalysis products.

In Fig. 1 the July zonal wind mean computed from the NSIPP-1 three-member

ensemble Control run is also shown. Given the uncertainties in the representation

of the AEJ among different data sets, some degree of resemblance suggests that

the NSIPP-1 is a reasonable simulation. The AEJ is clearly defined and well

separated from the low-level Harmattan flow. Its maximum is at about 650 hPa,

in agreement with the 8-year observation-based climatology (Burpee 1972), the

ERA-40, NCEP R2 and JRA-25 data. The maximum is somewhat weaker, being

16

about 7 m s-1 , and it is also slightly displaced to the north. The relative weakness

of the AEJ appears to be, in our model, a function of resolution. In experiments

with the same model run at higher resolution ( 1 ° x 1 °) the AEJ intensity is on the

order of 9 m s-1 (not shown), close to the ERA-40 results. The shallow west-

erly monsoonal flow and the TEJ are also well defined. As for the AEJ life-cycle,

the NSIPP simulation agrees well with the reanalyses with a northward shift of

about 60N in the AEJ position occurring between May and July, and a symmetric

southwad displacement, with decline in speed, in October (not shown). Moreover,

being the positions of the AEJ, of the monsoonal flow and of the TEJ well rep-

resented at the peak of the monsoon, there is confidence that the NSIPP model

is a valuable tool to investigate the forcings affecting the AEJ, even if the maxi-

mum speed of the AEJ as represented by the model is lower than the reanalyses.

The purpose of this work is to gain a deeper understanding of the mechanisms

controlling the basic AEJ structure and maintenance, and for this goal the NSIPP

representation of the AEJ structure appears satisfactory.

To further investigate some properties of the AEJ, the structure of the AEJ

with respect to the meridional horizontal shear of the zonal wind, is displayed in

Figure 2. This quantity, corresponding to the relative vorticity produced by the

zonal wind, is indicative of the part of the AEJ which may be more conducive

to barotropic instability to develop. All the 3 data sets show a strip of positive

values at about on the sourthern flank of the AEJ. The corresponding meridional

cross-sections localize the heigths and latitudes which appear, according to these

data sets, more likely to be barotropically unstable. All data sets agree that the

17

strongest cyclonic shear induced by the AEJ is at about 600 hPa and at about

10°N , but the ERA-40 display a more confined maximum. The cross-sections also

emphasize that in the ERA-40 the cyclonic shear appears to slope equatorward

with height, which is the opposite of what is observed in the JRA-25.

To further investigate the problem of barotropic instability in the AEJ, we plot

the quantity∂∂U

K(y) = ∂y [f (y) — ∂y ]

(4)

where f is the Coriolis parameter, and U is the zonal wind. K (Y) = 0 represents

the well-known necessary condition for barotropic instability to occur in response

to the zonal component of the flow. In our case, K (y) = 0 on the cyclonic side

of a flow indicates locations where barotropic instability is not impossible. This

corresponds to K (y) being zero to the north of positive values of K and to the

south of negative values of K (when the flow is easterly). In Figure3 K (y) is

plotted against the zonal wind. The 3 data sets differ substantially on this regard.

In particular the ECMWF builds a contiguous line with K=0 on the southern flank

of the AEJ, indicating that a strip of latitudes at about 10°N across the entire

continent is potentially unstable for cyclonic disturbances to grow at the expenses

of the mean flow. This contiguous feature is absent from the other data sets. Most

remarkable is the presence, again in the ERA-40, of a contiguous vertical line

of K (y) = 0 sloping slightly poleward with increasing latitude and decreasing

height, almost reaching the surface at about 121N . A full investigation of the AEJ

instabilities as AEWs-producing mechanisms should consider both baroclinic and

barotropic instability as done by Hsieh and Cook (2008). This however would be

18

beyond the purpose of this work, which is concerned on the AEJ structure and

maintenance, and not on AEW production.

Figure 4 shows the mean vegetation and the orography used in the Control

run. The vegetation follows a mosaic approach so that each gridbox is actually a

mosaic of vegetation types with strong subgrid variability. The colors in Figures 4

identify the predominant vegetation type in each gridbox. The subgrid variability

of land surface specifics cannot be easily represented in a map, but it should be

emphasized that each grid box consists of up to six different kinds of land prop-

erties, including albedoes, roughness, vegetation cover, etc. The orography of the

Control at this resolution is relatively smooth compared to reality, but the most

important topographic features of the African continent are represented: namely

the southern plateau, the Ethiopian highlands, the Atlas range, and the lower el-

evations of western Sahara. The effects on atmospheric dynamics resulting from

the suppression of these features, as we will show later, are actually very strong.

Controlling the evaporation and modifying the vegetation distribution has pow-

erful effects on the distribution of surface temperatures. Figure 5 shows the sur-

face temperatures produced by the experiments, and the changes with respect to

the Control. The uniform β that uses Sahelian values everywhere cools the Sahara

by more than 5°C and warms the area south of it by about 3 — 5°C. When Saha-

ran values of β are imposed, the meridional temperature gradient changes in the

opposite direction, with no change over the Sahara but a heating of up to 7 — 9°C

in the Sahel and to the south of it.

The uniform vegetation (Sahel type) imposed in the Experiment 3 leads to a

19

cooling of the southern part of the Sahara and a warming in the northern part. The

effect is opposite to Experiment 1, in which a constant 0 is imposed.

The comparison of Experiment 1 (uniform 0) and Experiment 3 (uniform veg-

etation) explains Experiment 4, in which both evaporation and vegetation proper-

ties are imposed to Sahelian values. In other words, the particular response of

surface temperature to the changes in vegetation and evaporation is nonlinear and

derives from the addition of two partly contrasting forcings.

Most interesting is the impact of the suppressed orography, which causes an

increase of surface temperature comparable to the one in which uniform values

of (Saharan) 0 are imposed: namely a heating of 7 — 9oC, with maximum on the

Ethiopian region. In addition to the temperature change induced by lower eleva-

tion, the increase is also connected with the reduction of topographic uplift which

is in nature imposed on the monsoonal flow by the Ethopian highland, and that is

being suppressed by the experiment. This causes a strong reduction in precipita-

tion and latent heat flux. The heating is stronger where the reduction of precipi-

tation is larger, and demonstrates the importance of the particular topography of

northern Africa in controlling the monsoonal flow.

Figure 6 shows the JAS precipitation and the difference from the Control in

all the experiments. The most striking result is that increasing 0 over the Sahara

(Experiment 1) does not produce large changes in precipitation while the opposite

change, imposing Saharan 0 everywhere, leads to a large reduction in precipitation

over the Sahel. Another interesting and somewhat counterintuitive result is that

prescribing uniform Sahelian type vegetation over Africa produces a much larger

20

impact than prescribing Sahelian-type 0, which leads to an impressive northward

shift of the ITCZ (Experiment 3).

Experiment 4, in which both evaporation and vegetation are prescribed, clearly

results in some nonlinear effects, in which the vegetation change (and presumably

the albedo) dominates the evaporability change.

Of great interest is the result of the orography experiment (Experiment 5)

which alone, without any modification to other surface forcings, can produce a

dramatic reduction of precipitation throughout the Sahel region. The orographic

effect, in fact, is comparable to the 0 effect in Experiment 2, where Saharan val-

ues of evaporability are imposed. In other words, the removal of the topography

of northern Africa can produce a reduction of precipitation almost equivalent to

the one obtained by forcing the evaporability to be virtually zero.

Figure 7 displays the corresponding JAS surface latent heat fluxes in the vari-

ous experiments, and their difference from the Control. Consistent with the above,

Sahelian and Saharan values of 0 alter the meridional gradients of latent heat (with

Sahelian 0 increasing latent heat over the Sahara and Saharan 0 reducing latent

heat south of the Sahara). Vegetation change and Sahelian 0 change act in the

same direction (increasing latent heat north of the Sahel, and decreasing it to the

south), but the impact scale of 0 is larger. This is reasonable because to prescribe

0 over the desert leads to very strong evaporation with consequent large latent

heat anomalies.

Figure 8 compares the meridional surface temperature gradients, a zonal cross-

section of meridional temperature gradients at 12°N and the temperature structure

21

in a meridional vertical cross section at 0° longitude, in the Control and vari-

ous Experiments. The displacement of precipitation induced by the Experiments

corresponds well to the displacement of the strongest surface temperature gradi-

ents. The zonal cross-section of meridional gradients is computed from 180°W to

180°E . However, changes with respect to the Control are noticeable, as expected,

only over the longitudes corresponding to the African continent, where the sur-

face conditions are altered in the experimental setup. Of particular significance

is the suppression of the meridional thermal gradient inversion, on the top of the

AEJ, between about 20°W and 60°E, at about 600 hPa. In the Control, merid-

ional temperature gradients are positive below 600 hPa, and vanish or become

slightly negative at about 500 hPa-600 hPa. In the uniform vegetation experiment

(Experiment 3) and uniform vegetation and uniform β experiment (Experiment 4)

the meridional temperature gradients, albeit smaller with height, remain always

positive. The meridional cross-sections of temperature at 0° longitude further re-

inforce the notion that the thermal structure of the atmospheres is dramatically

altered as a consequence of the Experiments, with evident changes in position and

amplitude of the thermal maxima.

In Fig. 9 the zonal wind at 600 hPa, and a meridional vertical cross section at

0° longitude are shown. The common feature between Experiments 1, 3 and 4 is

the thickening and strengthening of the low-level westerly monsoonal flow at 0°

longitude. The Harmatthan flow is not suppressed in any experiment, since being

simply driven by the always present summer thermal low over the Sahara.

In both Experiments 1 (with constant evaporability) and 3 (constant Sahelian

22

vegetation), the AEJ is still present, albeit displaced and with different intensity

(stronger and higher in Experiment 3, weaker and lower in Experiment 1).

It is in the Experiment 4, in which both vegetation and evaporation are simul-

taneously controlled, that the AEJ disappears. The relatively ‘moderate’ changes

in the boundary forcings performed in Experiments 1 and 3 are not sufficient to

suppress the AEJ, but only to change its intensity and position. The most drastic

Experiment 2, which imposes Saharan-like conditions (through constant 0) over

all Africa does suppress the jet, but it is very important to note that a less extreme

value of constant 0 as the one obtained from grassland is not sufficient to destroy

the jet. In other words, a ‘moderate’ control of evaporation (and thus soil mois-

ture gradients) alone is not sufficient to suppress the AEJ, but is the combination

of controlled evaporation and vegetation properties (inclusive of albedo) gradients

together which suppresses the Jet. This finding partly contradict Cook’s (1999)

conclusions: in that study an AEJ was produced by the use of realistic soil mois-

ture imposed over uniform albedo and the conclusion was made that only soil

moisture is relevant. In this present work it is shown, by separating the various

forcings, that the realistic vegetation types (inclusive of properties such as albedo)

can also produce an AEJ even if evaporation is controlled. Since Cook 1999 did

not produce an experiment with realistic albedo and prescribed evaporation, it is

suggested in this work that Cook’s (1999) conclusions should be revisited.

Moreover, it is important to notice that the experiment in which orography

is suppressed, but evaporation gradients and realistic vegetation are maintained,

leads to a different circulation with a weaker monsoonal flow and, most important,

23

no AEJ. In other words, the role of orography in controlling the AEJ is at least as

important as the role of combined vegetation and evaporation gradients. This is

another aspect of Cook’s (1999) conclusion that needs to be revisited: in that study

the orography was considered flat.

In Figure 10, similarly for what was done in Figure ??, the 600 zonal wind and

the corresponding meridional vertical sections at 0° longitude, together with the

meridional horizontal shear induced by the zonal wind, are shown to emphasize

the areas that are more ‘active’ from the point of view of barotropic instability

in the various experiments. It is important to realize that the areas with stronger

cyclonic vorticity are not necessarily controlled by the stronger AEJ, being also

partly controlled by the monsoonal flow. Cyclonic vorticity occurs on the south-

ern side of an easterly jet but also on the northern side of a westerly flow: in

both cases barotropic instability can develop. The Experiments seem to suggest

that what the AEJ controls is not so much the occurrence of instability, but the

confinement of it. For example, Experiment 4, which is remarkable for the ab-

sence of an AEJ, does have a broad area of cyclonic shear, mostly induced by the

strengthened monsoonal flow, and confined more to the north and to the east of

the Control. On the other hand the experiment with uniform 0 imposed to Sa-

haran values shows almost no area over northern Africa where cyclonic vorticity

dominates, confirming the overall absence of precipitation noted in Figure 5.

The changes in lower midtropospheric circulation associated with the Exper-

iments are summarized in Fig. 11, by comparing the Control and Experiments

geopotential and wind at 600 hPa and 925 hPa. The suppression of a strong east-

24

erly flow at about 15°N in the Experiment 4 is remarkable. 600 hPa wind speed

at 151N decreases from about 10m s-1 in the Control to few m s-1 in the Ex-

periment. This is consistent with the observed change in surface temperature, and

the strong weakening of meridional surface temperature gradients which occur in

Experiment 4 with respect to the control. Moreover, the southward penetration of

the predominantly low-level northerly flow from the eastern Mediterranean feed-

ing into the northern branch of the Hadley cell over Africa, is slightly reduced in

the Experiment 4, in which both vegetation properties and evaporation are con-

trolled.

In contrast, the orography experiment allows the flow from the Mediterranean

to directly collide with the monsoonal flow, essentially advecting dry air into the

ITCZ, consistent with the drastic reduction in precipitation and latent heat shown

in Figs. 6 and 7.

Being that the AEJ constrained by the thermal wind relationship, the recip-

rocal position of surface temperature gradients and zonal winds are emphasized

in Fig. 12 for the Control and Experiments 3 and 5, which displays meridional

surface temperature gradients against 600 hPa zonal wind. Experiments 3 and

5 are chosen as the two most extreme cases of meridional displacement of sur-

face temperature gradients: in Experiment 3 the maximum gradient is displaced

as far north as 20°N, whereas in Experiment 5 is displaced southward over west-

ern Africa, and almost to the Equator over Eastern Africa. While in Experiment

3 a feature resembling the AEJ can be identified in good correspondence to the

maximum temperature surface gradient, Experiment 5 shows that basically the

25

AEJ disappears merging within the predominatly easterly flow associated with

the ITCZ. This impressive change emphasizes the importance of orography in

creating meridional surface temperature gradients.

The overall findings indicate that the AEJ is a very fragile element of the Trop-

ical Meteorology, strongly constrained by boundary forcings, and that variations

in these constrains may lead to a completely different atmospheric regime.

6. Discussion and dynamical implications

This work starts with an analysis of the AEJ using the ERA-40 data and compar-

ing the AEJ representation with that in other state-of-the-art reanalysis data sets,

namely the JRA-25 and the NCEP R2. Despite some agreement among these rep-

resentations, and despite some consistency with pre-existing data sets regarding

the definition of the AEJ, it is found that some degree of ambiguity in terms of

AEJ shape and intensity still persists, indicating the need for a denser observa-

tional coverage over the Sahelian region.

A Control simulation and a set of five experiments are then performed using

the NSIPP-1 Atmospheric General Circulation Model. The Control is designed to

be a realistic representation of the atmosphere in which, among other forcings, re-

alistic land-surface properties, realistic orography and realistic SSTs are adopted.

The model has previously been demonstrated to satisfactorily reproduce climate

variability (Schubert et al. 2000; Giannini et al. 2003). The Control run produces

a reasonable circulation over Africa and represents well the AEJ. The five exper-

iments are designed to test the roles of various forcings, namely vegetation type,

26

soil moisture gradients (with two settings), vegetation type and soil moisture gra-

dients combined together, and orography. A sophisticated treatment of vegetation

is used, that determines a whole set of land surface properties including, but not

limited to, albedo.

The most important result is that the meridional thermal gradients associated

with the AEJ are primarily controlled by gradients of continental surface proper-

ties. In a GCM experiment in which both the evaporability and vegetation of the

land surface are prescribed to uniform, moderate values (i.e., those obtained in

a Control climate run over the Sahel region) over the entire African continent, a

general circulation is produced that completely suppresses the AEJ. Interestingly,

experiments in which vegetation and evaporability are prescribed separately both

result in a distinct AEJ, albeit one of somewhat different strength and position.

This result contrasts with the idealized experiments of Cook (1999) in which soil

moisture gradients alone controlled the AEJ. Indeed, a similar effect occurs in our

model when the soil moisture gradients are eliminated in an extreme experiment

in which we effectively set the evaporability to zero (i.e., to Saharan values) over

the entire continent. However, this experiment not only eliminates the AEJ but

also the African ITCZ and results in a very unrealistic general circulation over the

entire African monsoon region.

Among the various changes noted in the general circulation produced by Ex-

periment 4, in which both vegetation and evaporation are controlled, we empha-

size: a) the weakening of the Harmattan flow, b) the corresponding strengthen-

ing of the monsoonal flow and c) the overall increased zonality of the circu-

27

lation over Northern Africa. These changes are consistent with the Reale and

Dirmeyer (2002) findings on the effects of suppressing the interannual variabil-

ity of β. Those experiments were performed with the same methodology as in

KA00, but globally and on a different model. The Reale and Dirmeyer (2002)

study, based on a single realization, found a mid-tropospheric negative vorticity

anomaly over the Sahelian region suggestive of a northward shift of the ITCZ and

a weakening of the AEJ as a consequence of disabling globally the interannual

variability of β.

From this perspective, the results from this study, which are obtained with

a more localized change (only Africa) and with a 3-member ensemble, are much

more robust and support the idea that a combination of meridional gradients of soil

moisture and vegetation properties (including albedo) are crucial for maintaining

the AEJ.

The topography experiment (Experiment 5) is also relevant. Suppressing the

topography, of which a prominent feature are the Ethiopian Highlands, well repre-

sented in the model’s resolution (Fig. 4), leads to a strong increase in temperature

over East and equatorial Africa and a dramatic reduction of precipitation in the

eastern side of the African ITCZ. These two factors act in the same direction of

Experiment 4: namely the disappearance of meridional positive surface temper-

ature gradients at about 10° — 151N (Fig. 8). In other words, the combination

of vegetation properties and soil moisture gradients produces the same effects of

orography, particularly over eastern Africa.

Since the Control experiment includes all of these forcings (varying vegeta-

28

tion, varying evaporability and realistic topography), and since the suppression

of the first two leads to the disappearance of the AEJ as does the suppression of

the third, it can be safely argued that orography is just as important a player in

the maintenance of the AEJ, as the vegetation and evaporability. This finding was

not accessible to previous experiments that used more idealized simulations with

unrealistic orography.

Finally, another important aspect is the nonlinearity of the response to the com-

bined action of prescribing 0 and vegetation properties together. The prescription

of uniform Sahelian values of 0 over the entire continent has an effect on tem-

perature gradients that is opposite in sign to that of imposing uniform vegetation.

However, when both forcings are prescribed together in Experiment 4, their ef-

fects do not appear additive. This is further clarified in Fig. 13 where surface

latent heat flux, temperature and precipitation anomalies (computed as Experi-

ment minus Control) are displayed together for Experiments 1, 3, 4 and compared

with the simple sum of the anomalies computed from Experiments 1 and 3. When

constant 0 and constant vegetation act together in Experiment 4, their net result is

very different from the simple addition of the individual anomalies.

7. Conclusions

This work investigates the African Easterly Jet within the framework of a realis-

tic global model simulation. The current understanding of the AEJ is discussed

with the aid of reanalyses data sets, and five experiments are designed to assess

the importance of vegetation, soil moisture gradients and orography for forming

29

and maintaining the Jet. The results further strengthen the original intuition at the

basis of the Charney feedback (Charney et al. 1977) by confirming the sensitivity

to vegetation change of the Sahel. In particular, decrease in vegetation (corre-

sponding in part to a increase in albedo) over the Sahel leads to an altered surface

radiation balance which in turn leads to decreased evaporation and decreased pre-

cipitation.

However, the increased realism of vegetation modeling, which now includes

many other properties besides albedo, and the more sophisticated treatment of

evaporation introduce new insights and add substantial complexity to the problem.

Several steps have been taken in this direction by various authors. In particular, in

the process of further refining our understanding of the African meteorology, this

work aims to continue the research performed by Cook (1999).

Because of the use of a more realistic model with higher vertical and hori-

zontal resolution, and the use of multiple realization ensemble runs, this study

provides a more robust assessment of the mechanisms contributing to the AEJ.

The results presented here confirm that soil moisture is indeed an important forc-

ing, thus partly agreeing with previous findings by Cook (1999). However, the

set of experiments performed also suggest that not only evaporation gradients,

but a combination of vegetation properties and orography are also essential for

the maintenance of the mean climatological AEJ. In other words, according to

this study, it is the very particular combination of moisture gradients, vegetation

distribution and orography that produces the AEJ, not the soil moisture gradients

alone.

30

Finally, this work emphasizes the fragility of the African Easterly Jet, which

appears as a somewhat fortuitous element of tropical African meteorology, sub-

ject to fluctuations or even disappearance whenever one of the boundary forcings

is altered. In particular, there are several studies documenting the possible im-

pact of deforestation on regional climate: among these, changes in precipitation

and squall systems propagation on the Amazon (Da Silva et al., 2008); global

changes affecting Europe Asia and the United States (Findell et al. 2007) and

even teleconnections arising out of possible future deforestation over the Amazon

and central Africa (Avissar and Werth 2005). Most relevant is the study by Taylor

et al. (2002) which investigated the impact of land surface change on Sahelian

climate, demonstrating its sensitivity to small changes in albedo and leaf area in-

dex. It is therefore likely that change in vegetation over the Sahel may impact

precipitation, soil moisture gradients and consequently even the AEJ.

This aspect has some socio-economic implications, suggesting that land-surface

use over the Sahelian region should be carefully considered and that large-scale

dramatic changes in terms of deforestation and clearings to the south of the Jet,

over Equatorial Africa, may have repercussions on regional and larger-scale cli-

mate.

Acknowledgments This work was supported by the NASA Earth Science Enter-

prise’s Global Modeling and Analysis Program.

31

REFERENCES

Arakawa, A., and M. J. Suarez, 1983: Vertical differencing of the primitive

equations in sigma coordinates. Mon. Wea. Rev., 111, 34-45.

Asnani, G. C., 2005: Tropical Meteorology. Publisher: Indian Institute of

Tropical Meteorology, Pashan, Pune-411008, India. 3 vols.

Avissar, R., and D. Werth, 2005: Global Hydroclimatological Teleconnections

Resulting from Tropical Deforestation. J. Climate, 6, 134-145.

Bacmeister, J. T., P. J. Pegion, S. D. Schubert, and M. J. Suarez, 2000: Atlas of

seasonal means simulated by the NSIPP-1 atmospheric GCM. NASA God-

dard Space Flight Center, NASA Tech. Memo 104606, Vol. 17, 194pp.

,and M. J. Suarez, 2002: Wind stress simulations and the equatorial mo-

mentum budget in an AGCM. J. Atmos. Sci, 59, 3051-3073.

Burpee, R. W., 1972: The origin and structure of easterly waves in the lower

troposphere of north Africa. J. Atmos. Sci, 29, 77-90.

Charney, J., W. J. Quirk, S.-h. Chow, and J. Kornfield, 1977: A comparative

study of the effects of albedo change on drought in semi-arid regions.. J.

Atmos. Sci., 34, 1366-1385.

Chen, T.-C., and H. van Loon, 1987: Interannual variation of the Tropical

Easterly Jet. Mon. Wea. Rev., 115, 1739-1759.

Chen, T. H., and Coauthors, 1997: Cabauw experimental results from the

Project for Intercomparison of Land-Surface Parameterization Schemes.

J. Climate, 10, 11941215..

32

Chou, M.-D., and M. J. Suarez, 1994: An efficient thermal infrared radiation

parametrization for use in general circulation models. NASA Goddard

Space Flight Center, NASA Tech. Memo 104606, Vol 3, 85 pp.

,and—, 1999: A solar radiation parametrization for use in general circu-

lation models. NASA Goddard Space Flight Center, NASA Tech. Memo

104606, Vol 11, 40 pp.

Clark, D. B., C. M. Taylor, and A. J. Thorpe, 2004: Feedback between the

Land Surface and rainfall at convective lenght scales. J. Hydrometeor., 5,

625-639.

Cook, K., 1999: Generation of the African Easterly Jet and Its Role in Deter-

mining West African Precipitation. J. Climate, 14, 1165-1184.

,and E. K. Vizy, 2006: Coupled model simulations of the West African

Monsoon System: twentieth- and twenty-first-century simulations. J. Cli-

mate, 19, 3861-3703.

Da Silva, R. R., D. Werth, and R. Avissar, 2008: Regional Impacts of Fu-

ture Land-Cover Changes on the Amazon Basin Wet-Season Climate, J.

Climate, 21, 11531170.

Dhonneur, G. 1971: General circulation and types of weather over western

and central Africa. Annex IV, GATE Design, JOC, WMO, 22pp.

Douville, H., F. Chauvine, and H. Broqua, 2005: Influence of soil moisture

on the Asian and African Monsoons. Part I: Mean Monsoon and daily

precipitation. J. Climate, 14, 2381-2403.

33

Giannini, A., R. Saravanan, and P. Chang, 2003: Oceanic forcing of Sahel

rainfall on interannual to interdecadal time scales. Science, 302, 1027-

1030. DOI: 10.1126/science.1089357.

Findell, K. L., E. Shevliakova, P. C. D. Milly, and R. J. Stouffer, 2007: Mod-

eled Impact of Anthropogenic Land Cover Change on Climate. J. Climate,

20, 3621-3634.

Foland, C. K., Palmer, T. N., Parker, D. E. 1986: Sahel rainfall and worldwide

sea temperatures. Nature, 320, 602-607.

Gibson, J.K., P. Kallberg, S. Uppala, A. Hernandez, A. Nomura and E. Ser-

rano, 1997: ERA Description. ECMWF Re-Analysis Project Report Se-

ries, 1.

Hall N. M. J., G. N. Kiladis, and C. D. Thorncroft, 2006: Three-Dimensional

Sturcture and Cynamics of African Easterly Waves. Part II:Dynamical

Modes, J. Atmos. Sci., 63, 2231-2245.

Hamilton, R. A., and J. W. Archibold, 1945: Meteorology of Nigeria and

adiacent territories. Q. J. Roy. Met. Soc., 71, 231-264.

Hagos, S. M., and K. Cook, 2005: Influence of surface processes over Africa

on the Atlantic Marine ITCZ and South American Precipitation. J. Cli-

mate, 18, 4993-5010.

Houze, R. A., Jr., 1977: Structure and dynamics of a tropical squall-line sys-

tem observed during GATE. Mon. Wea. Rev., 105, 1540-1567.

Hsieh, J.-S., and K. Cook, 2005: Generation of African Easterly Wave distur-

34

bances: Relationship to the African Easterly Jet. Mon. Wea. Rev. 133,

1311-1327.

,and—, 2007: A study of the energetics of African Easterly Waves usign a

regional climate model. J. Atmos. Sci. 64, 421-440.

,and—, 2008: On the instability of the African Easterly Jet and the gen-

eration of African Waves: reversal of the potential vorticity gradient. J.

Atmos. Sci. 65, 2130-2151.

Kallberg, P., A. Simmons, S. Uppala, and M. Fuentes, 2004: The ERA-40

archive. ERA-40 Project Report Series, 17, 1-31. Available online at

http://www.ecmwf.int/publications/library/do/references/list/1 92

Kanamitsu M., W. Ebisuzaki, J. Woollen, S.-K. Yang, J. J. Hnilo, M. Fiorino,

and G. L. Potter, 2002: NCEPDOE AMIP-II Reanalysis (R-2). Bull. Amer.

Meteor. Soc., 83, 16311643.

Kiladis, G. N., C. D. Thorncroft, and N. M. J. Hall, 2006: Three-Dimensional

Sturcture and Cynamics of African Easterly Waves. Part I:Observations,

J. Atmos. Sci., 63, 2212-2230.

Kistler, R., E. Kalnay, W. Collins, S. Saha, G. White, J. Woollen, M. Chelliah,

W. Ebisuzaki, M. Kanamitsu, V. Kousky, H. van den Dool, R. Jenne, and

M. Fiorino, 2001: The NCEP-NCAR 50-year reanalysis: Monthly means

CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82, 247–267.

Koster, R. D., and M. J. Suarez, 1992: Modeling the land surface boundary

in climate models as a composite of independent vegetation stands. J.

35

Geophys. Res., 97, 26972715.

,and—, 1995: Relative contributions of land and ocean processes to pre-

cipitation variability. J. Geoph. Res., 100, 13,775-13,790.

,and—, 1996: Energy and water balance calculations in the Mosaic LSM.

NASA Goddard Space Flight Center, NASA Tech. Memo. 104606, Vol.

9., 59 pp.

,—, and M. Heiser, 2000: Variance and Predictability of Precipitation at

Seasonal-to-Interannual Timescales. J. Hydrometeor., 1, 26-46.

Louis, J., M. Tiedtke, and J. Geleyn, 1982: A short history of the PBL parametriza-

tion at ECMWF. Proc. ECMWF Workshop on Planetary Boundary Layer

Parametrization, Reading, United Kingdom, ECMWF, 59-80.

Moorthi, S., and M. J. Suarez, 1992: Relaxed Arakawa-Schubert: A parametriza-

tion of moist convection for general circulation models. Mon. Wea. Rev.,

120, 978-1002.

Nicholson, S. E., 1997: An analysis of the ENSO signal in the Tropical At-

lantic and western Indian Ocean. Int. J. Climatology., 17, 345-375.

,and J. Kim, 1997: The relationship of the ENSO to African Rainfall. Int.

J. Climatology., 17, 117-135.

Onogi, K., J. Tsutsui, H. Koide, M. Sakamoto, S. Kobayashi, H. Hatsushika,

T. Matsumoto, N. Yamazaki, H. Kamahori, K. Takahashi, S. Kadokura, K.

Wada, K. Kato, R. Oyama, T. Ose, N. Mannoji and R. Taira, 2007: The

JRA-25 Reanalysis. J. Meteor. Soc. Japan, 85, 369-432.

36

Parker, D. J., A. Fink, S. Janicot, J.-B. Ngamini, M. Douglas, E. Afiesimama,

A. Agusti-Panared, A. Beljaars, F. Dide, A. Diedhiou, T. Lebel, J. Polcher,

J.-L. Redelsperger, C. Thorncroft, and G. Ato Wilson, 2008: The AMMA

radiosonde program and its implications for the future of atmospheric

monitoring over Africa. Bull. Am. Meteor. Soc., 89, 1015-1027.

Pegion, P. J., Schubert, and M. J. Suarez, 2000: An assessment of the pre-

dictability of northern winter seasonal means with the NSIPP 1 AGCM.

NASA Goddard Space Flight Center, NASA Tech. Memo 104606, Vol.

18, 110 pp.

Reale, O., and P. Dirmeyer, 2002: Modeling the effect of land-surface evapo-

ration variability on precipitation variability. Part I: General response. J.

Hydrometeor., 3, 433-450.

—,—, and A. Schlosser, 2002: Modeling the effect of land-surface evapora-

tion variability on precipitation variability. Part II: Time- and space-scale

structure. J. Hydrometeor., 3, 451-466.

Schubert, S. D., M. J. Suarez, P. J. Pegion, M. A. Kistler, and A. Kumer, 2002:

Predictability of zonal means during boreal summer. J. Climate, 15, 420-

434.

Semazzi, F. H. M., B. Burns, N.-H. Lin, and J.-K. Schemm, 1996: A GCM

study of the teleconnections between the continental climate of Africa and

Global Sea Surface Temperature anomalies. 2002: Predictability of zonal

means during boreal summer. J. Climate, 9, 2480-2497.

37

Simmsons, A., and J.K. Gibson, 2000: The ERA-40 project plan. ERA-40

Project Report Series, 1, 1-63. Available online at

http://www.ecmwf.int/publications/library/do/references/list/1 92

Suarez M. J., and L. L. Takacs, 1995: Documentation of the Arries-GEOS

dynamical core: version 2. NASA Goddard Space Flight Center, NASA

Tech. Memo. 104606, Vol. 5, 45 pp.

Taylor, C. M., E. F. Lambin, N. Stephenne, R. J. Hardling, and R. L. H. Essey,

2002: The Influence of Land Use Change on Climate in the Sahel. J.

Climate, 15, 36153629.

Tompkins, A. M., C. Cardinali, J.-J. Morcrette, and M. Rodwell, 2005: In-

fluence of aerosol climatology on forecasts of the African Easterly Jet.

Geoph. Res. Letter., 32, L10801, doi:10.1029/2004GL022189.

Thorncroft, C. D., and M. Blackburn, 1999: Maintenance of the African east-

erly jet Q. J. Roy. Met. Soc., 125, 763-786.

,D. J. Parker, R. R. Burton, M. Diop, J. H. Ayers, H. Bariat, S. Devereau,

A. Diongue, R. Dumelow, D. R. Kindred, N. M. Price, M. Saloum, C.

M. Tayor, and A. M. Tompkins, 2003: The JET 2000 Project. Aircraft

observations of the African Easterly Jet and African Easterly Waves. Bull.

Am. Meteor. Soc., 84, 337-351.

Redelsperger, J.-L., C. D. Thorncroft, A. Diedhiou, T. Lebel, D. J. Parker,

and J. Polcher, 2006: African Monsoon Multidisciplinary Analysis: An

international research project and field campaign. Bull. Am. Meteor. Soc.,

38

87, 1739-1746.

Tucker, C., Dregne, H. D., and W. W. Newcomb, 1999: Expansiopn and con-

traction of the Sahara desert from 1980 to 1990. Science, 253, 299-300.

DOI: 10. 1 126/science.253.5017.299.

Uppala, S. M., P. W. Kallberg, A. J. Simmons, U. Andrae, V. D. Bechtold et al.,

2005: The ERA-40 re-analysis. Q. J. R. Meteorol. Soc., 131, 2961-3012.

Xue, Y., and J. Shukla, 1993: The Influence of Land Surface Properties on

Sahel Climate. Part 1: Desertification J. Climate, 6, 2232-2245.

Wu, C. M.-L., S. Schubert, M. J. Suarez, and P. J. Pegion, 2006: Seasonality

and Meridional Propagation of the MJO. J. Climate, 19, 1901-1921.

Wood, E. F., and Coauthors, 1998: The Project for the Intercomparison of

Land-Surface Parameterization Schemes (PILPS) phase 2(c) Red-Arkansas

River basin experiment. 1: Experiment description and summary inter-

comparisons. Global Planet. Change, 19,115-135.

Vizy, E. K, and K. H. Cook, 2001: Mechanisms by which Gulf of Guinea and

Eastern North Atlantic Sea Surface Temperature anomalies can influence

African rainfall. J. Climate, 14, 795-821.

Zipser, E. J., and C. Gautier, 1978: Mesoscale Events within a GATE tropical

depression Mon. Wea. Rev., 106, 789-805.

Zeng, N. and J. D. Neelin, 2000: The role of vegetation-climate interaction

and interannual variability in shaping the African Savanna. J. Climate, 13,

2665-2670.

39

List of Tables

1 The experiment settings are described. The first column shows

the role of albedo and vegetation type. The second column shows

the role of the evaporation efficiency 0. The third column refers

to the orography. . . . . . . . . . . . . . . . . . . . . . . . . . . 41

40

Vegetation Evaporation Efficiency OrographyControl Varying Varying RealisticEXP 1 Varying Prescribed (Sahel conditions) RealisticEXP 2 Varying Prescribed (Sahara conditions) RealisticEXP 3 Prescribed Varyiing RealisticEXP 4 Prescribed Prescribed RealisticEXP 5 Varying Varyiing Prescribed

Table 1: The experiment settings are described. The first column shows the role of

albedo and vegetation type. The second column shows the role of the evaporation

efficiency 0. The third column refers to the orography.

41

List of Figures

1 July zonal wind (m s-1 , contours every 1m s-1 , 0 omitted, solid-positive,

dashed-negative) climatology (1980-2001) based on ERA-40, NCEP R2,

JRA-25 data, and NSIPP-1 AGCM three ensemble runs: 600 hPa (left

column) and cross section at 0° longitude (right column). . . . . . . . 45

2 July zonal wind (m s-1 , contours every 1m s-1 , 0 omitted, solid-positive,

dashed-negative) climatology (1980-2001) based on ERA-40, NCEP R2,

and JRA-25 data, and meridional horizontal shear of the zonal wind at

600 hPa (left column, contours at +/ — 2,1.5, 1,.5.1 x 10 -6 s-1). Cross

section at 0° longitude (right column). . . . . . . . . . . . . . . . . . 46

3 July zonal wind (m s-1 , contours every 1m s-1 , 0 omitted, solid-positive,

dashed-negative) climatology (1980-2001) based on ERA-40, NCEP R2,

and JRA-25 data, and K(y) at 600 hPa (left column, contours at +/ —

2,1, .5 x 10-11 m-1 s-1). Cross section at 0° longitude (right column). . 47

4 Map of predominant vegetation types. The actual vegeatation in the

model is much more complex, since up to six vegetation types can be ob-

served within each gridbox, following a ‘mosaic’ approach (left panel).

Map of model topography (right panel). . . . . . . . . . . . . . . . . 48

5 JAS surface temperature (K) computed for the Control and Experiments

1, 2, 3, 4, 5 (left column, upper color bar). Right column: differences

from the control (lower color bar) . . . . . . . . . . . . . . . . . . . . 49

42

6 JAS precipitation (mm d-1 ) computed for the Control and Experiments

1, 2, 3, 4, 5 (left column, upper color bar). Right column: differences

from the control (lower color bar) . . . . . . . . . . . . . . . . . . . . 50

7 JAS latent heat fluxes (Wm-2) computed for the Control and Experi-

ments 1, 2, 3, 4, 5 (left column, upper color bar). Right column: differ-

ences from the control (lower color bar) . . . . . . . . . . . . . . . . . 51

8 1980-2001 meridional surface temperature gradient ( K deg-1) com-

puted for the Control and Experiments 1, 2, 3, 4, 5 (left column, up-

per color bar). Zonal vertical cross-sections of meridional temperature

gradient (K deg-1) at 12°N (central column, middle color bar). Merid-

ional vertical cross-sections of temperature ( K) at 0° longitude (right

column, lower color bar) . . . . . . . . . . . . . . . . . . . . . . . . 52

9 July zonal wind (m s-1) computed for the Control and Experiments 1,

2, 3, 4, 5 (left column). Right column: meridional vertical cross sections

of zonal wind at 0° longitude. . . . . . . . . . . . . . . . . . . . . . 53

10 July zonal wind (m s-1) computed for the Control and Experiments 1,

2, 3, 4, 5 (left column) and meridional horizontal shear of the zonal wind

at 600 hPa (left column, contours at +/ — 2,1.5,1, .5.1 × 10-6 s-1, solid-

positive, dashed-negative). Right column: meridional vertical cross sec-

tions of the same quantities at 0° longitude. . . . . . . . . . . . . . . 54

11 Geopotential height (hm) and wind (m s-1 ) at 600 hPa (left column, up-

per color bar) and 925 hPa (right column, lower color bar) in the Control

and Experiment runs. . . . . . . . . . . . . . . . . . . . . . . . . . 55

43

12 600 hPa zonal wind (m s-1 , contours every 1m s- 1 , 0 omitted, solid-

positive, dashed-negative) in the Control and Experiments 3 and 5, com-

pared with meridional surface temperature gradients (shaded K deg-1.) 56

13 Anomalies (computed as the difference Experiment minus Control) of

surface latent heat fluxes (first column), surface temperature (second

column) and precipitation (third column) for Experiment 1 (first row),

Experiment 3 (second row), Experiment 4 (third row). The last row

shows the anomaly sums of Experiment 1 plus Experiment 3.. . . . . . 57

44

Figure 1: July zonal wind (m s-1 , contours every 1m s- 1 , 0 omitted, solid-positive,dashed-negative) climatology (1980-2001) based on ERA-40, NCEP R2, JRA-25 data,and NSIPP-1 AGCM three ensemble runs: 600 hPa (left column) and cross section at 0°longitude (right column).

45

Figure 2: July zonal wind (m s-1 , contours every 1m s- 1 , 0 omitted, solid-positive,dashed-negative) climatology (1980-2001) based on ERA-40, NCEP R2, and JRA-25data, and meridional horizontal shear of the zonal wind at 600 hPa (left column, contoursat +/ — 2, 1.5, 1,.5.1 x 10-6s-1). Cross section at 0° longitude (right column).

46

Figure 3: July zonal wind (m s-1 , contours every 1m s- 1 , 0 omitted, solid-positive,dashed-negative) climatology (1980-2001) based on ERA-40, NCEP R2, and JRA-25data, and K(y) at 600 hPa (left column, contours at +/ — 2,1, .5 x 10- 11 m- 1 s- 1 ).Cross section at 0° longitude (right column).

47

Figure 4: Map of predominant vegetation types. The actual vegeatation in the model ismuch more complex, since up to six vegetation types can be observed within each gridbox,following a ‘mosaic’ approach (left panel). Map of model topography (right panel).

48

Figure 5: JAS surface temperature (K) computed for the Control and Experiments 1, 2,3, 4, 5 (left column, upper color bar). Right column: differences from the control (lowercolor bar).

49

Figure 6: JAS precipitation (mm d-1 ) computed for the Control and Experiments 1, 2,3, 4, 5 (left column, upper color bar). Right column: differences from the control (lowercolor bar).

50

Figure 7: JAS latent heat fluxes (Wm-2) computed for the Control and Experiments1, 2, 3, 4, 5 (left column, upper color bar). Right column: differences from the control(lower color bar).

51

Figure 8: 1980-2001 meridional surface temperature gradient ( K deg— ') computed forthe Control and Experiments 1, 2, 3, 4, 5 (left column, upper color bar). Zonal verticalcross-sections of meridional temperature gradient (K deg— ') at 12°N (central column,middle color bar). Meridional vertical cross-sections of temperature ( K) at 0° longitude(right column, lower color bar).

52

Figure 9: July zonal wind (m s-1 ) computed for the Control and Experiments 1, 2, 3,4, 5 (left column). Right column: meridional vertical cross sections of zonal wind at 0°longitude.

53

Figure 10: July zonal wind (m s-1 ) computed for the Control and Experiments 1, 2,3, 4, 5 (left column) and meridional horizontal shear of the zonal wind at 600 hPa (leftcolumn, contours at +/- 2,1.5,1, .5.1 × 10-6 s-1, solid-positive, dashed-negative). Rightcolumn: meridional vertical cross sections of the same quantities at 0° longitude.

54

Figure 11: Geopotential height (hm) and wind (m s-1 ) at 600 hPa (left column, uppercolor bar) and 925 hPa (right column, lower color bar) in the Control and Experimentruns.

55

Figure 12: 600 hPa zonal wind (m s-1 , contours every 1m s- 1 , 0 omitted, solid-positive,dashed-negative) in the Control and Experiments 3 and 5, compared with meridional sur-face temperature gradients (shaded K deg-1.)

56

Figure 13: Anomalies (computed as the difference Experiment minus Control) of surfacelatent heat fluxes (first column), surface temperature (second column) and precipitation(third column) for Experiment 1 (first row), Experiment 3 (second row), Experiment 4(third row). The last row shows the anomaly sums of Experiment 1 plus Experiment 3.

57